Intro

Exploring price patterns by day of week for several major cryptocurrencies.

Data

Get data:

  • Start with BTC and ETH
  • Using Cdn$ since…I make and spend my money in Canada.
## [1] "BTC-CAD"
##      Index             BTC-CAD.Open      BTC-CAD.High      BTC-CAD.Low     
##  Min.   :2014-09-17   Min.   :  211.2   Min.   :  253.7   Min.   :  205.1  
##  1st Qu.:2016-08-09   1st Qu.:  802.5   1st Qu.:  806.7   1st Qu.:  799.8  
##  Median :2018-07-02   Median : 8436.4   Median : 8538.8   Median : 8288.0  
##  Mean   :2018-07-02   Mean   :15192.8   Mean   :15588.6   Mean   :14758.1  
##  3rd Qu.:2020-05-24   3rd Qu.:14194.6   3rd Qu.:14525.9   3rd Qu.:13887.5  
##  Max.   :2022-04-16   Max.   :84050.8   Max.   :85358.0   Max.   :82605.9  
##                       NA's   :1         NA's   :1         NA's   :1        
##  BTC-CAD.Close    
##  Min.   :  212.7  
##  1st Qu.:  803.8  
##  Median : 8436.3  
##  Mean   :15208.9  
##  3rd Qu.:14240.6  
##  Max.   :84072.1  
##  NA's   :1

Focus on recent data

  • Get last 2 years
  • Check overall pattern
  • Identify days of week
## 'data.frame':    731 obs. of  9 variables:
##  $ BTC.CAD.Open    : num  9376 9992 9930 10166 10098 ...
##  $ BTC.CAD.High    : num  10016 10034 10178 10193 10184 ...
##  $ BTC.CAD.Low     : num  9260 9894 9925 10034 9658 ...
##  $ BTC.CAD.Close   : num  9992 9935 10161 10102 9721 ...
##  $ BTC.CAD.Volume  : num  6.57e+10 4.55e+10 4.54e+10 4.40e+10 5.33e+10 ...
##  $ BTC.CAD.Adjusted: num  9992 9935 10161 10102 9721 ...
##  $ date            : Date, format: "2020-04-16" "2020-04-17" ...
##  $ day             : chr  "Thu" "Fri" "Sat" "Sun" ...
##  $ week_of         : Date, format: "2020-04-12" "2020-04-12" ...

Focusing on closing price on the assumption that if there is a pattern it will show up in closing price.

No strong, obvious pattern over the period. Monday has lowest median, Wednesday has highest, but LOTS of spread in the data, with no reliable pattern.

This data is over a 2 year period where there is a lot of variance in the price over time that may blur the weekly patterns. Could still be that there is a consistent pattern within weeks.

Let’s look at week-by-week trends by day:

Certainly some weeks with upward trend through the week (most obviously near end of 2021), but not exactly a consistent pattern to rely on across this date range.

Maybe focus on the past year, when price is in higher bracket.

Look at some individual day of week comparisons

  • Start with Sun - Sat

If anything, looks like tendency to go down throughout the week from Sun to Sat? Let’s take a closer look:

Basically a wash:

  • 50% chance of being either above or below 0.12% return on the week, although farther outliers on the negative side.
  • Ave. return on the week essentially 0 (-0.012%).

Check other days

Open vs Close

Another comparison is checking if some days tend to close higher than open. This could lead to a strategy of buying early and selling late on days that tend to have upward trend.

Check Linear Model

## 
## Call:
## lm(formula = BTC.CAD.Close ~ day, data = btc_rec_df)
## 
## Residuals:
##    Min     1Q Median     3Q    Max 
## -19626  -8183  -1368   6722  26570 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 57509.473   1519.032  37.859   <2e-16 ***
## dayMon         -7.563   2148.236  -0.004    0.997    
## dayTue       -100.962   2148.236  -0.047    0.963    
## dayWed        -52.653   2148.236  -0.025    0.980    
## dayThu       -498.690   2148.236  -0.232    0.817    
## dayFri       -529.423   2148.236  -0.246    0.805    
## daySat        -68.817   2148.236  -0.032    0.974    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 10950 on 357 degrees of freedom
##   (1 observation deleted due to missingness)
## Multiple R-squared:  0.0003891,  Adjusted R-squared:  -0.01641 
## F-statistic: 0.02316 on 6 and 357 DF,  p-value: 0.9999